the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Oct 2024
Technical Note: Benchmark time-temperature paths provide a shared framework for evaluating and communicating thermochronologic data interpretation
Abstract. We present a set of six time-temperature (tT) histories, called benchmark paths, that can be used as a shared framework for evaluating the sensitivity of a thermochronologic system to the variables inherent in the interpretation of thermochronologic data (e.g., kinetics models, mineral compositions or geometries, etc.). These benchmark paths span 100 Myr, include monotonic and nonmonotonic histories that represent plausible geologic scenarios, and have a range of cooling rates through different chronometer partial-retention/annealing temperatures. Here, we demonstrate their utility by presenting a method for tuning these paths to 11 different kinetics models for the apatite (U-Th-Sm)/He (n=5), apatite fission-track (n=2), and zircon (U-Th)/He (n=4) systems. These tuned tT paths provide a practical comparison of the kinetics models for each system and the data patterns they predict, thereby offering anyone performing thermal history analysis the ability to consider how their choice of kinetics model may impact their data interpretation. The adoption of benchmark paths for evaluating kinetics models and other variables provides a practical way for the thermochronology community to evaluate and communicate the decision making processes that are inherent in thermochronologic modeling and data interpretation.
Competing interests: All authors are Coordinators for the GChron Special Issue “Technical notes on modelling thermochronologic data” to which this technical note is submitted.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(724 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on gchron-2024-20', Anonymous Referee #1, 17 Oct 2024
Goddard et al. present six time-temperature (t-T) paths that represent "plausible geologic scenarios" for evaluating the sensitivity of different thermochronologic systems, ultimately to be used in evaluating and communicating data interpretations.
The following questions could be asked:
1. Is there a need/is it a good idea to standardize t-T benchmark curves?
2. Is the proposed selection of t-T benchmarks fit for purpose?
3. Should the authors make this decision on behalf of the thermochronology community?These questions are addressed in order below.
1. I do not think there is a need, nor is it a good idea, to "standardize" t-T curves. I’m not sure of the exact reason for doing this either? Many if not all of the t-T curves are those published [or variations thereof] in Wolf et al. (1998), which were also discussed in the recent Murray et al. (2022) HeFTy modeling summary paper (their Figures 2 and 3). So it isn’t clear to me what is new in the manuscript other than the proposal to include this information more formally/separately in the GChron 'special issue’ on interpretation/modeling. The authors do expand the applied thermochronometers beyond apatite (U-Th)/He and fission track (in Wolf et al.) to include zircon (U-Th)/He, but this does not justify a new contribution. Beyond this, the predicted data from such "benchmark" t-T paths can be generated by anyone wanting to examine those histories on their own and refer to the original Wolf et al. work, just like the author's did in their 2022 Geosphere publication.As stated in Lines 63–70 of the manuscript: "We demonstrate the utility of our proposed benchmark paths by using them to illustrate the different temperature sensitivities of three low-temperature thermochronometers (AHe, AFT, ZHe), and then, within each system, how kinetics models also require different temperatures to produce the same age. This is useful because although experimentally-derived kinetics models provide the foundation for the interpretation of thermochronologic data, it can be difficult to develop a practical understanding of if or how choosing one kinetics model over another might impact one’s thermal history model results. This is critical for both project design and data interpretation." (italics added here)...and stated in Lines 217–225, Section 5: "A vision for the application of benchmark paths. Here, we demonstrate how a suite of benchmark tT paths can be designed to leverage the temperature sensitivity of a particular low-temperature thermochronometer and then tuned to specific kinetics models. We propose that the six benchmark paths we use in this work can provide a practical tool for the thermochronology community to use in a variety of contexts, including comparing kinetics models and predicting data patterns that arise from variable mineral compositions or geometries. This ‘design-then-tune’ approach is not meant to identify a single ‘best’ kinetics model for a particular system but to quantify and visualize how kinetics models predict different tT conditions and data patterns." (italics added here)This is simply forward modeling a t-T path and extracting the predicted data—one of the most fundamental things we do. What is being tuned? I think it is generally understood that kinetic models will produce different predictions, or so-called 'data patterns'. Kinetic models do not strictly predict "different tT conditions" but in this case instead yield model ages under a specified t-T condition. Furthermore, all of the identified kinetic models for (U-Th)/He and fission track have flaws and varying degrees of uncertainty, with many of the uncertainties and/or assumptions being different between models for the same system (e.g., different AFT analysts and compositional data/quantities considered in Ketcham et al. 1999 v. 2007 kinetic models). The same path producing slightly different age predictions is inherent to such differences in kinetic models. More often than not, kinetic models are a progression in understanding and improved calibration. Therefore, the answer may be that the most recent kinetic model is the preferable one, unless there is a specific reason for choosing otherwise (i.e., different fundamental radiation damage assumptions between Guenthner v. Ginster in zircon). For example, no one is going to try to publish t-T models using the original Wolf '96/Farley '00 helium diffusion kinetics that do not account for radiation damage. In terms of how choice of kinetics impacts t-T model results, there a number of papers that already show these sorts of effects (e.g., Guenthner 2021). In light of this, showing the effects again here for "benchmark" paths does not seem particularly useful given the near infinite number of t-T histories that could be also considered a benchmark by another modeler. I could see this sort of exercise being useful for training students or for use as a module in an university course. However, it would be better implemented as an activity and not reading material.2. I would say that the answer to this question is debatable. There are many t-T paths that could considered representative benchmarks, and that could be argued in different ways and mostly depends on perspective and context. Due to those reasons alone, one could perhaps argue that “standardization” is necessary, but I would actually think the opposite. Why were these particular paths chosen, are they truly representative of end member geological environments, or is it simply because they exist? The t-T paths presented in the manuscript were conjured in the original Wolf et al. publication but that doesn't mean that those t-T paths are special in any meaningful way.3. If something like this were to be done it would require broader input as to the reasons why this is necessary and what the very specific usage of such standardized paths would actually be. I do not think this sort of standardization is useful or beneficial on behalf of the community and could actually generate confusion going forward if someone shows a certain t-T path as an example and they are instead “told” to use one of these “benchmark” histories. Again, a similar exercise can be done by anyone modeling 'real world' data to get an understanding of the sensitivity across different thermochronometric mineral systems with respect to a proposal history.References:Farley, K. A.: Helium diffusion from apatite: General behavior as illustrated by Durango fluorapatite, J. Geophys. Res. Solid Earth, 105, 2903–2914, https://doi.org/10.1029/1999jb900348, 2000.Guenthner, W. R.: Implementation of an Alpha Damage Annealing Model for Zircon (U‐Th)/He Thermochronology With Comparison to a Zircon Fission Track Annealing Model, Geochemistry, Geophys. Geosystems, 22, https://doi.org/10.1029/2019GC008757, 2021.Ketcham, R. A., Donelick, R. A., and Carlson, W. D.: Variability of apatite fission-track annealing kinetics: III. Extrapolation to geological time scales, Am. Mineral., 84, 1235–1255, https://doi.org/10.2138/am-1999-0903, 1999.Ketcham, R. A., Carter, A., Donelick, R. A., Barbarand, J., and Hurford, A. J.: Improved modeling of fission-track annealing in apatite, Am. Mineral., 92, 799–810, https://doi.org/10.2138/am.2007.2281, 2007.Murray, K. E., Goddard, A. L. S., Abbey, A. L., and Wildman, M.: Thermal history modeling techniques and interpretation strategies: Applications using HeFTy, Geosphere, 18, 1622–1642, https://doi.org/10.1130/GES02500.1, 2022.Wolf, R. A., Farley, K. A., and Silver, L. T.: Helium diffusion and low-temperature thermochronometry of apatite, Geochim. Cosmochim. Acta, 60, 4231–4240, https://doi.org/10.1016/S0016-7037(96)00192-5, 1996.Wolf, R. A., Farley, K. A., and Kass, D. M.: Modeling of the temperature sensitivity of the apatite (U-Th)/He thermochronometer, Chem. Geol., 148, 105–114, https://doi.org/10.1016/S0009-2541(98)00024-2, 1998.Citation: https://doi.org/10.5194/gchron-2024-20-RC1 -
AC1: 'Reply on RC1', Andrea Stevens Goddard, 17 Dec 2024
In our reply, we copied referee comments for reference and have replied to each of those comments in italics below.
Anonymous Referee #1
Goddard et al. present six time-temperature (t-T) paths that represent "plausible geologic scenarios" for evaluating the sensitivity of different thermochronologic systems, ultimately to be used in evaluating and communicating data interpretations.
The following questions could be asked:
- Is there a need/is it a good idea to standardize t-T benchmark curves?
We address this broad question following the specific points in the referee’s comments below.
- I do not think there is a need, nor is it a good idea, to "standardize" t-T curves. I’m not sure of the exact reason for doing this either? Many if not all of the t-T curves are those published [or variations thereof] in Wolf et al. (1998), which were also discussed in the recent Murray et al. (2022) HeFTy modeling summary paper (their Figures 2 and 3). So it isn’t clear to me what is new in the manuscript other than the proposal to include this information more formally/separately in the GChron 'special issue’ on interpretation/modeling. The authors do expand the applied thermochronometers beyond apatite (U-Th)/He and fission track (in Wolf et al.) to include zircon (U-Th)/He, but this does not justify a new contribution. Beyond this, the predicted data from such "benchmark" t-T paths can be generated by anyone wanting to examine those histories on their own and refer to the original Wolf et al. work, just like the author's did in their 2022 Geosphere publication.
Similar paths have been shown before, but the use and application of benchmark paths is new. Below we summarize the new intellectual contributions of this Technical Note.
Intellectual contributions include:
- A proposed formalization of using a suite of six classic thermal histories, which we call “benchmark paths,” to which we argue provides a needed and simple shared framework for comparing and quantifying differences in kinetics models, tT history features, mineral compositions/geometries, and other variables critical to the interpretation of thermochronologic data.
- A novel tuning method for these benchmark paths, which permits practical comparisons of the tT history differences that arise from such variables. We demonstrate this tuning approach using different chronometer kinetics models.
As stated in Lines 63–70 of the manuscript: "We demonstrate the utility of our proposed benchmark paths by using them to illustrate the different temperature sensitivities of three low-temperature thermochronometers (AHe, AFT, ZHe), and then, within each system, how kinetics models also require different temperatures to produce the same age. This is useful because although experimentally-derived kinetics models provide the foundation for the interpretation of thermochronologic data, it can be difficult to develop a practical understanding of if or how choosing one kinetics model over another might impact one’s thermal history model results. This is critical for both project design and data interpretation."
...and stated in Lines 217–225, Section 5: "A vision for the application of benchmark paths. Here, we demonstrate how a suite of benchmark tT paths can be designed to leverage the temperature sensitivity of a particular low-temperature thermochronometer and then tuned to specific kinetics models. We propose that the six benchmark paths we use in this work can provide a practical tool for the thermochronology community to use in a variety of contexts, including comparing kinetics models and predicting data patterns that arise from variable mineral compositions or geometries. This ‘design-then-tune’ approach is not meant to identify a single ‘best’ kinetics model for a particular system but to quantify and visualize how kinetics models predict different tT conditions and data patterns."
This is simply forward modeling a t-T path and extracting the predicted data—one of the most fundamental things we do.
We agree that forward modeling is a fundamental tool for thermochronologists. However, many people who generate and interpret thermochronologic data are not trained as thermochronologists, and moreover, forward modeling is not a method that is typically clearly described in research publications. In part, that is because it is often part of the data exploration process that does not become formalized in a final manuscript. For this reason, not all scientists generating and interpreting thermochronology data have learned this fundamental skill. We can attest from our extensive experiences teaching thermochronologic data interpretation and modeling to students and scientists from many institutions and backgrounds that forward modeling is not widely understood or practiced. The purpose of this Technical note (and this Special Issue of Technical Notes) is to make the steps and decision making that is a critical part of thermochronologic modeling and interpretation explicit. We recognize that this contribution has a different style, content, and purpose than a typical research publication, but we believe that it is nonetheless valuable to communicate widely through formal review and publication.
What is being tuned? I think it is generally understood that kinetic models will produce different predictions, or so-called 'data patterns'. Kinetic models do not strictly predict "different tT conditions" but in this case instead yield model ages under a specified t-T condition.
In a revised publication,we will provide a more explicit definition of our tuning approach and try to make it more clear that the tuning took careful thought and effort - it is not as straight forward as it looks. Our tuning method can be used in a few different ways, as described in the text, but the purpose is to quantify practical differences that arise from some variable of interest—in our example application, choosing different kinetics models during thermal history analysis.
For example, in this technical note, we illustrate several simple points that should be obvious to all thermochronologists but are, in our experience as reviewers, research advisors, and instructors, commonly not understood by geoscientists who have not had the opportunity to work in-depth with thermochronologic data. We use our tuning approach to illustrate how in simple rapid cooling scenarios, the choice of kinetics model does not change the maximum temperature and cooling rate needed to produce the same single AHe age from a grain with a fixed composition and size. Then, we use other benchmark paths to demonstrate this is not true for all styles of tT histories. For example, for histories with a reheating event, in order to produce that same AHe age, different kinetics models require different maximum temperatures, and thus cooling rates. The differences among the “tuned” tT paths thus provide a practical quantification and visualization of the differences among kinetics models. Again, although we agree that “it is generally understood that kinetic models will produce different predictions, or so-called 'data patterns" we argue that this statement only applies to thermochronologists who spend a lot of time doing thermal history analysis. Additionally, we respectfully disagree with the reviewer’s suggestion that the practical consequences of these kinetics differences—and that one can use forward modeling to think about these differences—are obvious from the current literature.
This, however, is just one example of tuning. We also point out in the text that tuning can be used to reflect the consequences of other thermal history model inputs (e.g., etching protocols, grain geometry, mineral chemistry). Importantly, the tuning approach provides a framework for a reader to build upon, following their own questions.
Furthermore, all of the identified kinetic models for (U-Th)/He and fission track have flaws and varying degrees of uncertainty, with many of the uncertainties and/or assumptions being different between models for the same system (e.g., different AFT analysts and compositional data/quantities considered in Ketcham et al. 1999 v. 2007 kinetic models). The same path producing slightly different age predictions is inherent to such differences in kinetic models.
While it is generally understood that kinetics models produce different data patterns, there is no practice for either quantifying or developing an intuition for those differences. Benchmark paths address this formally. Our tuning approach also provides an alternative method for conceptualizing these differences.
More often than not, kinetic models are a progression in understanding and improved calibration. Therefore, the answer may be that the most recent kinetic model is the preferable one, unless there is a specific reason for choosing otherwise (i.e., different fundamental radiation damage assumptions between Guenthner v. Ginster in zircon).
We agree that kinetics models can be a progression of understanding; however it is actually quite common to have parallel models in use. In addition to the example of the zircon (U-Th)/He system (Guenthner et al., 2013; Ginster et al., 2019), there are three parallel models in use for the apatite (U-Th)/He system. The Gautheron et al., (2009) and Willett et al., (2017) kinetics models for the apatite (U-Th)/He system both α-recoil damage to define radiation damage, whereas the Flowers et al., (2009) kinetics model for the apatite (U-Th)/He system uses fission track damage as a proxy for radiation damage. Benchmark paths provide a way to quantify how these different kinetics models - all viable candidates to apply to data interpretation - predict/interpret data distribution differently in an intuitive way that can help modelers understand the uncertainties in their data sets that are related to our understanding of radiation damage.
For example, no one is going to try to publish t-T models using the original Wolf '96/Farley '00 helium diffusion kinetics that do not account for radiation damage.
We have two responses to this point:
- Some of these early kinetics models are still in use. For example, codes/programs that calculate thermokinematic exhumation spatially often use simpler kinetics models because of the complexity of the computations. Pecube, a three-dimensional heat transport equation that models a crustal/lithospheric block undergoing uplift and surface erosion, by default uses the Farley et al., (2000) and Reiners et al., (2004) kinetics models for the apatite and zircon (U-Th)/He systems, respectively. A recent code published in GChron to calculate steady-state vertical exhumation rates, age2exhume, also uses these kinetics models. We want to emphasize that we are not critiquing the use of these kinetics models in either of these programs. In fact, the exercises in this manuscript using benchmark paths demonstrates that these kinetics models may be appropriate for relatively fast and monotonic exhumation. However, the important point is that the users of these programs are able to consider how interpreting data using the kinetics models in these programs may impact the interpretations.
- Most recent thermal history models do not apply these early kinetics models. However, we still argue that using benchmark paths to contextualize the differences between these kinetics models and the subsequent generation of kinetics models is critical for evaluating legacy data sets. This is particularly important for recently trained thermochronologists who have not followed the development of these models in real time.
In terms of how choice of kinetics impacts t-T model results, there a number of papers that already show these sorts of effects (e.g., Guenthner 2021). In light of this, showing the effects again here for "benchmark" paths does not seem particularly useful given the near infinite number of t-T histories that could be also considered a benchmark by another modeler. I could see this sort of exercise being useful for training students or for use as a module in an university course. However, it would be better implemented as an activity and not reading material.
We agree with the point that this is a useful exercise for training. This is a strong argument for publication. Not all students advancing their knowledge of thermochronology techniques and modeling abilities have access to the same training. And not all folks advancing their knowledge of thermochronology techniques and modeling abilities are students, or are scientists who have access to someone who could design and teach such a course. It is absolutely critical with the expanded use of thermochronology data and modeling that we properly train individuals in understanding the behaviors of thermochronologic systems. Published demonstrations provide equitable access that is critical to building a well trained thermochronology community. This, in fact, is the purpose of this Special Issue which arose out of a community call for these kinds of contributions at the past two Thermo meetings to provide contributions to train and advance the field of thermochronology in an accessible and equitable way.
- Is the proposed selection of t-T benchmarks fit for purpose?
I would say that the answer to this question is debatable. There are many t-T paths that could considered representative benchmarks, and that could be argued in different ways and mostly depends on perspective and context. Due to those reasons alone, one could perhaps argue that “standardization” is necessary, but I would actually think the opposite. Why were these particular paths chosen, are they truly representative of end member geological environments, or is it simply because they exist? The t-T paths presented in the manuscript were conjured in the original Wolf et al. publication but that doesn't mean that those t-T paths are special in any meaningful way.
Benchmark paths are not the exhaustive solution to exploring the sensitivity of each thermochronometric system to differences in kinetics models, tT history features, mineral compositions/geometries, and other variables critical to the interpretation of thermochronologic data. However, benchmark paths offer a new, additional tool to systematically explore these sensitivities. They provide a community baseline that facilitates practical demonstrations and training, and - like all scientific contributions - can be built upon in the future. But only if the foundation is built in the first place.
- Should the authors make this decision on behalf of the thermochronology community?
If something like this were to be done it would require broader input as to the reasons why this is necessary and what the very specific usage of such standardized paths would actually be. I do not think this sort of standardization is useful or beneficial on behalf of the community and could actually generate confusion going forward if someone shows a certain t-T path as an example and they are instead “told” to use one of these “benchmark” histories.
We agree that it is important to have community buy-in for benchmark paths. We would be open to hearing the reviewer’s ideas about a vision for that expansion, but we did not identify any clear vision that could be built upon in this text.
As members of the thermochronology community we have made a concerted effort to formally and informally gather input on these ideas through teaching courses on modeling at the Thermo conferences and having explicit discussions with many in the Thermochronology community. While we know that no selection of benchmark paths will be exhaustive, we have made a thoughtful and thorough effort for community input on these selections. We also think that, like all published scientific work, it is important to introduce a contribution that can be built on. This text is intended to begin a formal and public conversation about how we use and understand thermochronologic systems. We welcome further contributions as this idea evolves.
We have explicitly addressed the selection of benchmark paths in the text stating that benchmark paths are not intended to exclude any other paths that provide an important demonstration of a kinetics model or other sensitivity/behavior of a thermochronologic system. To the contrary, these are very important! However the importance of these specific t-T paths does not usurp the value of an additional tool, benchmark paths. We include the text below that is used to make this point:
“These individualized tT histories remain a fundamental contribution because they demonstrated behaviors distinctive to a particular kinetics model and the rocks these models were first applied to. Our benchmark paths complement these contributions by providing a universal reference frame that can be used to compare these kinetics models.”
Again, a similar exercise can be done by anyone modeling 'real world' data to get an understanding of the sensitivity across different thermochronometric mineral systems with respect to a proposal history.
We agree that any individual can model sensitivity of real world data. The challenge is that with the growing user base of thermochronology, many people have not been trained to conduct sensitivity analysis. As we state in the text, benchmark paths can be used for a variety of applications because they are designed to highlight the sensitivity of each thermochronometric system to differences in kinetics models, tT history features, mineral compositions/geometries, and other variables critical to the interpretation of thermochronologic data. Introducing benchmark paths is a practical foundation for sensitivity analysis that is a tool for understanding the sensitivity of real world data.
Citation: https://doi.org/10.5194/gchron-2024-20-AC1 -
AC3: 'Statement on Goals of the Manuscript', Andrea Stevens Goddard, 17 Dec 2024
This manuscript was designed specifically for the Technical Note format and goals as described by GChron. It is not a research paper, but proposes how a new tool, benchmark paths, can be used to better understand and quantify the impacts of the decisions inherent in modeling and interpreting thermochronologic data. As part of our contribution we describe a process of tuning benchmark paths which permits practical, quantitative comparisons of the tT history differences that arise from modifying variables including, but not limited to, kinetics models, tT history features, mineral compositions/geometries. We demonstrate this tuning approach using different chronometer kinetics models.
This contribution provides formal guidance on modeling skills that are critical to robust data interpretations, but typically not formalized in research papers either in the main text or supplementary text. Peer-reviewed publication provides wide and equitable access to learning thermochronologic data analysis and interpretation techniques. The content is aligned with both the solicitation for Technical Notes by GChron and the Special Issue to which it was submitted, “Technical notes on modelling thermochronologic data”. Moreover, it is written to address a community need for formalized documentation and training in modeling techniques that has been part of a recurring community conversation at International Meetings (e.g. International Conference on Thermochronology, 2021 & 2023). A key audience for this publication is individuals within and outside the thermochronology community interested in growing their technical skills in modeling and data interpretation.
Citation: https://doi.org/10.5194/gchron-2024-20-AC3
-
AC1: 'Reply on RC1', Andrea Stevens Goddard, 17 Dec 2024
-
RC2: 'Comment on gchron-2024-20', Anonymous Referee #2, 18 Oct 2024
Why do we need new benchmarks? I don’t think this short paper has made a strong case. Tellingly, in the discussion (lines 227-232) authors note that ..”a single suite of tuned paths cannot capture all complexities of these systems”……. Very true so why are these new benchmarks fit for purpose?
We know models are imperfect and will continue to be improved upon. Yes, it is important to test and compare different models to determine their efficacy but this needs to be done in depth exploring the full range of predictions for a wide range of geological timescales, compositions and thermal history complexity. This study is a mere snapshot.
The greatest challenge with models is the extrapolation from laboratory to geological timescales. We can only know if a model works through comparison with geological data and this is why researchers were motivated to compare different (U-Th)/He and FT models with data from borehole samples with a well constrained independent geological/thermal history which is what Naeser, (1979, 1980, 1981); Green et al., (1989) and later, House et al., (2002) attempted to do.
The discussion highlights some of the many limitations/uncertainties of (U-Th)/He and FT models that undermine the case for benchmark paths because how do you know your benchmark is a sensible reference point given the dispersion of predictions present in other models. Perhaps your benchmarks are outliers.
A final point regarding experimental design. Why choose a single composition for both AFT and AHe? – Are they representative of the most common types of apatite? The choice of using DPAR is questionable in that a DPAR of 2.05 µm can be found in apatites with between 0.02 wt% Chlorine and up to 0.8 wt% chlorine. Model outcomes would very different for some thermal histories has chlorine been used instead of DPAR.
Citation: https://doi.org/10.5194/gchron-2024-20-RC2 -
AC2: 'Reply on RC2', Andrea Stevens Goddard, 17 Dec 2024
In our reply, we copied referee comments for reference and have replied to each of those comments in italics below.
Anonymous Referee #2
Why do we need new benchmarks? I don’t think this short paper has made a strong case. Tellingly, in the discussion (lines 227-232) authors note that ..”a single suite of tuned paths cannot capture all complexities of these systems”……. Very true so why are these new benchmarks fit for purpose?
We know models are imperfect and will continue to be improved upon. Yes, it is important to test and compare different models to determine their efficacy but this needs to be done in depth exploring the full range of predictions for a wide range of geological timescales, compositions and thermal history complexity. This study is a mere snapshot.
We want to point out that the intention of this Technical Note is not to determine the efficacy of kinematics models. Our intention is to develop a new tool, benchmark paths, to explore how different inputs in thermal history modeling should be considered in data interpretations. As an example, we use benchmark paths to explore the sensitivity of kinetics models to different types of thermal histories.
We agree that this is a snapshot, but importantly it introduces a new tool for comparing different models that we anticipate will be incorporated into texts introducing new kinetics models and also lays the foundations for the personal training and practice of thermochronologists to quantify differences among kinetics models themselves. We intentionally describe methods of tuning benchmark paths to explore these sensitivities so that this topic can be expanded formally or informally. A single, exhaustive work on this topic would be impractical and inefficient.
The greatest challenge with models is the extrapolation from laboratory to geological timescales. We can only know if a model works through comparison with geological data and this is why researchers were motivated to compare different (U-Th)/He and FT models with data from borehole samples with a well constrained independent geological/thermal history which is what Naeser, (1979, 1980, 1981); Green et al., (1989) and later, House et al., (2002) attempted to do.
We agree. The purpose of benchmark paths is not to establish the efficacy of a particular model, but to quantify (in an intuitive way) differences among models. Geological tests of models are incredibly valuable, but beyond the purpose and scope of this Technical Note.
The discussion highlights some of the many limitations/uncertainties of (U-Th)/He and FT models that undermine the case for benchmark paths because how do you know your benchmark is a sensible reference point given the dispersion of predictions present in other models. Perhaps your benchmarks are outliers.
Whether a benchmark path is a sensible reference point will depend upon the geologic problem. The goal of this paper was to provide a strategic subset of benchmark paths that help modelers get a quantitative grasp of how thermochronologic data can be sensitive to differences in kinetics models, tT history features, mineral compositions/geometries, and other variables. In some cases benchmark paths may provide a sufficient understanding for a modeler to move forward. In other scenarios, a modeler may need to expand using different paths to explore sensitivities unique to their particular study. Importantly, this manuscript (1) formalizes the importance of exploring these questions, (2) provides a new tool, benchmark paths, which can be used by thermochronologists to explore uncertainties particularly relevant to their own data set, and (3) provides examples of how to use and tune benchmark paths.
A final point regarding experimental design. Why choose a single composition for both AFT and AHe? – Are they representative of the most common types of apatite? The choice of using DPAR is questionable in that a DPAR of 2.05 µm can be found in apatites with between 0.02 wt% Chlorine and up to 0.8 wt% chlorine. Model outcomes would very different for some thermal histories has chlorine been used instead of DPAR.
We agree! This is an additional sensitivity that can be explored, but was beyond the scope of this manuscript. As the coordinators of this Special Issue we anticipate that there are authors preparing demonstrations highlighting this exact point which we think is a very practical demonstration, particularly relevant for this Special Issue.
Citation: https://doi.org/10.5194/gchron-2024-20-AC2
-
AC2: 'Reply on RC2', Andrea Stevens Goddard, 17 Dec 2024
Status: closed
-
RC1: 'Comment on gchron-2024-20', Anonymous Referee #1, 17 Oct 2024
Goddard et al. present six time-temperature (t-T) paths that represent "plausible geologic scenarios" for evaluating the sensitivity of different thermochronologic systems, ultimately to be used in evaluating and communicating data interpretations.
The following questions could be asked:
1. Is there a need/is it a good idea to standardize t-T benchmark curves?
2. Is the proposed selection of t-T benchmarks fit for purpose?
3. Should the authors make this decision on behalf of the thermochronology community?These questions are addressed in order below.
1. I do not think there is a need, nor is it a good idea, to "standardize" t-T curves. I’m not sure of the exact reason for doing this either? Many if not all of the t-T curves are those published [or variations thereof] in Wolf et al. (1998), which were also discussed in the recent Murray et al. (2022) HeFTy modeling summary paper (their Figures 2 and 3). So it isn’t clear to me what is new in the manuscript other than the proposal to include this information more formally/separately in the GChron 'special issue’ on interpretation/modeling. The authors do expand the applied thermochronometers beyond apatite (U-Th)/He and fission track (in Wolf et al.) to include zircon (U-Th)/He, but this does not justify a new contribution. Beyond this, the predicted data from such "benchmark" t-T paths can be generated by anyone wanting to examine those histories on their own and refer to the original Wolf et al. work, just like the author's did in their 2022 Geosphere publication.As stated in Lines 63–70 of the manuscript: "We demonstrate the utility of our proposed benchmark paths by using them to illustrate the different temperature sensitivities of three low-temperature thermochronometers (AHe, AFT, ZHe), and then, within each system, how kinetics models also require different temperatures to produce the same age. This is useful because although experimentally-derived kinetics models provide the foundation for the interpretation of thermochronologic data, it can be difficult to develop a practical understanding of if or how choosing one kinetics model over another might impact one’s thermal history model results. This is critical for both project design and data interpretation." (italics added here)...and stated in Lines 217–225, Section 5: "A vision for the application of benchmark paths. Here, we demonstrate how a suite of benchmark tT paths can be designed to leverage the temperature sensitivity of a particular low-temperature thermochronometer and then tuned to specific kinetics models. We propose that the six benchmark paths we use in this work can provide a practical tool for the thermochronology community to use in a variety of contexts, including comparing kinetics models and predicting data patterns that arise from variable mineral compositions or geometries. This ‘design-then-tune’ approach is not meant to identify a single ‘best’ kinetics model for a particular system but to quantify and visualize how kinetics models predict different tT conditions and data patterns." (italics added here)This is simply forward modeling a t-T path and extracting the predicted data—one of the most fundamental things we do. What is being tuned? I think it is generally understood that kinetic models will produce different predictions, or so-called 'data patterns'. Kinetic models do not strictly predict "different tT conditions" but in this case instead yield model ages under a specified t-T condition. Furthermore, all of the identified kinetic models for (U-Th)/He and fission track have flaws and varying degrees of uncertainty, with many of the uncertainties and/or assumptions being different between models for the same system (e.g., different AFT analysts and compositional data/quantities considered in Ketcham et al. 1999 v. 2007 kinetic models). The same path producing slightly different age predictions is inherent to such differences in kinetic models. More often than not, kinetic models are a progression in understanding and improved calibration. Therefore, the answer may be that the most recent kinetic model is the preferable one, unless there is a specific reason for choosing otherwise (i.e., different fundamental radiation damage assumptions between Guenthner v. Ginster in zircon). For example, no one is going to try to publish t-T models using the original Wolf '96/Farley '00 helium diffusion kinetics that do not account for radiation damage. In terms of how choice of kinetics impacts t-T model results, there a number of papers that already show these sorts of effects (e.g., Guenthner 2021). In light of this, showing the effects again here for "benchmark" paths does not seem particularly useful given the near infinite number of t-T histories that could be also considered a benchmark by another modeler. I could see this sort of exercise being useful for training students or for use as a module in an university course. However, it would be better implemented as an activity and not reading material.2. I would say that the answer to this question is debatable. There are many t-T paths that could considered representative benchmarks, and that could be argued in different ways and mostly depends on perspective and context. Due to those reasons alone, one could perhaps argue that “standardization” is necessary, but I would actually think the opposite. Why were these particular paths chosen, are they truly representative of end member geological environments, or is it simply because they exist? The t-T paths presented in the manuscript were conjured in the original Wolf et al. publication but that doesn't mean that those t-T paths are special in any meaningful way.3. If something like this were to be done it would require broader input as to the reasons why this is necessary and what the very specific usage of such standardized paths would actually be. I do not think this sort of standardization is useful or beneficial on behalf of the community and could actually generate confusion going forward if someone shows a certain t-T path as an example and they are instead “told” to use one of these “benchmark” histories. Again, a similar exercise can be done by anyone modeling 'real world' data to get an understanding of the sensitivity across different thermochronometric mineral systems with respect to a proposal history.References:Farley, K. A.: Helium diffusion from apatite: General behavior as illustrated by Durango fluorapatite, J. Geophys. Res. Solid Earth, 105, 2903–2914, https://doi.org/10.1029/1999jb900348, 2000.Guenthner, W. R.: Implementation of an Alpha Damage Annealing Model for Zircon (U‐Th)/He Thermochronology With Comparison to a Zircon Fission Track Annealing Model, Geochemistry, Geophys. Geosystems, 22, https://doi.org/10.1029/2019GC008757, 2021.Ketcham, R. A., Donelick, R. A., and Carlson, W. D.: Variability of apatite fission-track annealing kinetics: III. Extrapolation to geological time scales, Am. Mineral., 84, 1235–1255, https://doi.org/10.2138/am-1999-0903, 1999.Ketcham, R. A., Carter, A., Donelick, R. A., Barbarand, J., and Hurford, A. J.: Improved modeling of fission-track annealing in apatite, Am. Mineral., 92, 799–810, https://doi.org/10.2138/am.2007.2281, 2007.Murray, K. E., Goddard, A. L. S., Abbey, A. L., and Wildman, M.: Thermal history modeling techniques and interpretation strategies: Applications using HeFTy, Geosphere, 18, 1622–1642, https://doi.org/10.1130/GES02500.1, 2022.Wolf, R. A., Farley, K. A., and Silver, L. T.: Helium diffusion and low-temperature thermochronometry of apatite, Geochim. Cosmochim. Acta, 60, 4231–4240, https://doi.org/10.1016/S0016-7037(96)00192-5, 1996.Wolf, R. A., Farley, K. A., and Kass, D. M.: Modeling of the temperature sensitivity of the apatite (U-Th)/He thermochronometer, Chem. Geol., 148, 105–114, https://doi.org/10.1016/S0009-2541(98)00024-2, 1998.Citation: https://doi.org/10.5194/gchron-2024-20-RC1 -
AC1: 'Reply on RC1', Andrea Stevens Goddard, 17 Dec 2024
In our reply, we copied referee comments for reference and have replied to each of those comments in italics below.
Anonymous Referee #1
Goddard et al. present six time-temperature (t-T) paths that represent "plausible geologic scenarios" for evaluating the sensitivity of different thermochronologic systems, ultimately to be used in evaluating and communicating data interpretations.
The following questions could be asked:
- Is there a need/is it a good idea to standardize t-T benchmark curves?
We address this broad question following the specific points in the referee’s comments below.
- I do not think there is a need, nor is it a good idea, to "standardize" t-T curves. I’m not sure of the exact reason for doing this either? Many if not all of the t-T curves are those published [or variations thereof] in Wolf et al. (1998), which were also discussed in the recent Murray et al. (2022) HeFTy modeling summary paper (their Figures 2 and 3). So it isn’t clear to me what is new in the manuscript other than the proposal to include this information more formally/separately in the GChron 'special issue’ on interpretation/modeling. The authors do expand the applied thermochronometers beyond apatite (U-Th)/He and fission track (in Wolf et al.) to include zircon (U-Th)/He, but this does not justify a new contribution. Beyond this, the predicted data from such "benchmark" t-T paths can be generated by anyone wanting to examine those histories on their own and refer to the original Wolf et al. work, just like the author's did in their 2022 Geosphere publication.
Similar paths have been shown before, but the use and application of benchmark paths is new. Below we summarize the new intellectual contributions of this Technical Note.
Intellectual contributions include:
- A proposed formalization of using a suite of six classic thermal histories, which we call “benchmark paths,” to which we argue provides a needed and simple shared framework for comparing and quantifying differences in kinetics models, tT history features, mineral compositions/geometries, and other variables critical to the interpretation of thermochronologic data.
- A novel tuning method for these benchmark paths, which permits practical comparisons of the tT history differences that arise from such variables. We demonstrate this tuning approach using different chronometer kinetics models.
As stated in Lines 63–70 of the manuscript: "We demonstrate the utility of our proposed benchmark paths by using them to illustrate the different temperature sensitivities of three low-temperature thermochronometers (AHe, AFT, ZHe), and then, within each system, how kinetics models also require different temperatures to produce the same age. This is useful because although experimentally-derived kinetics models provide the foundation for the interpretation of thermochronologic data, it can be difficult to develop a practical understanding of if or how choosing one kinetics model over another might impact one’s thermal history model results. This is critical for both project design and data interpretation."
...and stated in Lines 217–225, Section 5: "A vision for the application of benchmark paths. Here, we demonstrate how a suite of benchmark tT paths can be designed to leverage the temperature sensitivity of a particular low-temperature thermochronometer and then tuned to specific kinetics models. We propose that the six benchmark paths we use in this work can provide a practical tool for the thermochronology community to use in a variety of contexts, including comparing kinetics models and predicting data patterns that arise from variable mineral compositions or geometries. This ‘design-then-tune’ approach is not meant to identify a single ‘best’ kinetics model for a particular system but to quantify and visualize how kinetics models predict different tT conditions and data patterns."
This is simply forward modeling a t-T path and extracting the predicted data—one of the most fundamental things we do.
We agree that forward modeling is a fundamental tool for thermochronologists. However, many people who generate and interpret thermochronologic data are not trained as thermochronologists, and moreover, forward modeling is not a method that is typically clearly described in research publications. In part, that is because it is often part of the data exploration process that does not become formalized in a final manuscript. For this reason, not all scientists generating and interpreting thermochronology data have learned this fundamental skill. We can attest from our extensive experiences teaching thermochronologic data interpretation and modeling to students and scientists from many institutions and backgrounds that forward modeling is not widely understood or practiced. The purpose of this Technical note (and this Special Issue of Technical Notes) is to make the steps and decision making that is a critical part of thermochronologic modeling and interpretation explicit. We recognize that this contribution has a different style, content, and purpose than a typical research publication, but we believe that it is nonetheless valuable to communicate widely through formal review and publication.
What is being tuned? I think it is generally understood that kinetic models will produce different predictions, or so-called 'data patterns'. Kinetic models do not strictly predict "different tT conditions" but in this case instead yield model ages under a specified t-T condition.
In a revised publication,we will provide a more explicit definition of our tuning approach and try to make it more clear that the tuning took careful thought and effort - it is not as straight forward as it looks. Our tuning method can be used in a few different ways, as described in the text, but the purpose is to quantify practical differences that arise from some variable of interest—in our example application, choosing different kinetics models during thermal history analysis.
For example, in this technical note, we illustrate several simple points that should be obvious to all thermochronologists but are, in our experience as reviewers, research advisors, and instructors, commonly not understood by geoscientists who have not had the opportunity to work in-depth with thermochronologic data. We use our tuning approach to illustrate how in simple rapid cooling scenarios, the choice of kinetics model does not change the maximum temperature and cooling rate needed to produce the same single AHe age from a grain with a fixed composition and size. Then, we use other benchmark paths to demonstrate this is not true for all styles of tT histories. For example, for histories with a reheating event, in order to produce that same AHe age, different kinetics models require different maximum temperatures, and thus cooling rates. The differences among the “tuned” tT paths thus provide a practical quantification and visualization of the differences among kinetics models. Again, although we agree that “it is generally understood that kinetic models will produce different predictions, or so-called 'data patterns" we argue that this statement only applies to thermochronologists who spend a lot of time doing thermal history analysis. Additionally, we respectfully disagree with the reviewer’s suggestion that the practical consequences of these kinetics differences—and that one can use forward modeling to think about these differences—are obvious from the current literature.
This, however, is just one example of tuning. We also point out in the text that tuning can be used to reflect the consequences of other thermal history model inputs (e.g., etching protocols, grain geometry, mineral chemistry). Importantly, the tuning approach provides a framework for a reader to build upon, following their own questions.
Furthermore, all of the identified kinetic models for (U-Th)/He and fission track have flaws and varying degrees of uncertainty, with many of the uncertainties and/or assumptions being different between models for the same system (e.g., different AFT analysts and compositional data/quantities considered in Ketcham et al. 1999 v. 2007 kinetic models). The same path producing slightly different age predictions is inherent to such differences in kinetic models.
While it is generally understood that kinetics models produce different data patterns, there is no practice for either quantifying or developing an intuition for those differences. Benchmark paths address this formally. Our tuning approach also provides an alternative method for conceptualizing these differences.
More often than not, kinetic models are a progression in understanding and improved calibration. Therefore, the answer may be that the most recent kinetic model is the preferable one, unless there is a specific reason for choosing otherwise (i.e., different fundamental radiation damage assumptions between Guenthner v. Ginster in zircon).
We agree that kinetics models can be a progression of understanding; however it is actually quite common to have parallel models in use. In addition to the example of the zircon (U-Th)/He system (Guenthner et al., 2013; Ginster et al., 2019), there are three parallel models in use for the apatite (U-Th)/He system. The Gautheron et al., (2009) and Willett et al., (2017) kinetics models for the apatite (U-Th)/He system both α-recoil damage to define radiation damage, whereas the Flowers et al., (2009) kinetics model for the apatite (U-Th)/He system uses fission track damage as a proxy for radiation damage. Benchmark paths provide a way to quantify how these different kinetics models - all viable candidates to apply to data interpretation - predict/interpret data distribution differently in an intuitive way that can help modelers understand the uncertainties in their data sets that are related to our understanding of radiation damage.
For example, no one is going to try to publish t-T models using the original Wolf '96/Farley '00 helium diffusion kinetics that do not account for radiation damage.
We have two responses to this point:
- Some of these early kinetics models are still in use. For example, codes/programs that calculate thermokinematic exhumation spatially often use simpler kinetics models because of the complexity of the computations. Pecube, a three-dimensional heat transport equation that models a crustal/lithospheric block undergoing uplift and surface erosion, by default uses the Farley et al., (2000) and Reiners et al., (2004) kinetics models for the apatite and zircon (U-Th)/He systems, respectively. A recent code published in GChron to calculate steady-state vertical exhumation rates, age2exhume, also uses these kinetics models. We want to emphasize that we are not critiquing the use of these kinetics models in either of these programs. In fact, the exercises in this manuscript using benchmark paths demonstrates that these kinetics models may be appropriate for relatively fast and monotonic exhumation. However, the important point is that the users of these programs are able to consider how interpreting data using the kinetics models in these programs may impact the interpretations.
- Most recent thermal history models do not apply these early kinetics models. However, we still argue that using benchmark paths to contextualize the differences between these kinetics models and the subsequent generation of kinetics models is critical for evaluating legacy data sets. This is particularly important for recently trained thermochronologists who have not followed the development of these models in real time.
In terms of how choice of kinetics impacts t-T model results, there a number of papers that already show these sorts of effects (e.g., Guenthner 2021). In light of this, showing the effects again here for "benchmark" paths does not seem particularly useful given the near infinite number of t-T histories that could be also considered a benchmark by another modeler. I could see this sort of exercise being useful for training students or for use as a module in an university course. However, it would be better implemented as an activity and not reading material.
We agree with the point that this is a useful exercise for training. This is a strong argument for publication. Not all students advancing their knowledge of thermochronology techniques and modeling abilities have access to the same training. And not all folks advancing their knowledge of thermochronology techniques and modeling abilities are students, or are scientists who have access to someone who could design and teach such a course. It is absolutely critical with the expanded use of thermochronology data and modeling that we properly train individuals in understanding the behaviors of thermochronologic systems. Published demonstrations provide equitable access that is critical to building a well trained thermochronology community. This, in fact, is the purpose of this Special Issue which arose out of a community call for these kinds of contributions at the past two Thermo meetings to provide contributions to train and advance the field of thermochronology in an accessible and equitable way.
- Is the proposed selection of t-T benchmarks fit for purpose?
I would say that the answer to this question is debatable. There are many t-T paths that could considered representative benchmarks, and that could be argued in different ways and mostly depends on perspective and context. Due to those reasons alone, one could perhaps argue that “standardization” is necessary, but I would actually think the opposite. Why were these particular paths chosen, are they truly representative of end member geological environments, or is it simply because they exist? The t-T paths presented in the manuscript were conjured in the original Wolf et al. publication but that doesn't mean that those t-T paths are special in any meaningful way.
Benchmark paths are not the exhaustive solution to exploring the sensitivity of each thermochronometric system to differences in kinetics models, tT history features, mineral compositions/geometries, and other variables critical to the interpretation of thermochronologic data. However, benchmark paths offer a new, additional tool to systematically explore these sensitivities. They provide a community baseline that facilitates practical demonstrations and training, and - like all scientific contributions - can be built upon in the future. But only if the foundation is built in the first place.
- Should the authors make this decision on behalf of the thermochronology community?
If something like this were to be done it would require broader input as to the reasons why this is necessary and what the very specific usage of such standardized paths would actually be. I do not think this sort of standardization is useful or beneficial on behalf of the community and could actually generate confusion going forward if someone shows a certain t-T path as an example and they are instead “told” to use one of these “benchmark” histories.
We agree that it is important to have community buy-in for benchmark paths. We would be open to hearing the reviewer’s ideas about a vision for that expansion, but we did not identify any clear vision that could be built upon in this text.
As members of the thermochronology community we have made a concerted effort to formally and informally gather input on these ideas through teaching courses on modeling at the Thermo conferences and having explicit discussions with many in the Thermochronology community. While we know that no selection of benchmark paths will be exhaustive, we have made a thoughtful and thorough effort for community input on these selections. We also think that, like all published scientific work, it is important to introduce a contribution that can be built on. This text is intended to begin a formal and public conversation about how we use and understand thermochronologic systems. We welcome further contributions as this idea evolves.
We have explicitly addressed the selection of benchmark paths in the text stating that benchmark paths are not intended to exclude any other paths that provide an important demonstration of a kinetics model or other sensitivity/behavior of a thermochronologic system. To the contrary, these are very important! However the importance of these specific t-T paths does not usurp the value of an additional tool, benchmark paths. We include the text below that is used to make this point:
“These individualized tT histories remain a fundamental contribution because they demonstrated behaviors distinctive to a particular kinetics model and the rocks these models were first applied to. Our benchmark paths complement these contributions by providing a universal reference frame that can be used to compare these kinetics models.”
Again, a similar exercise can be done by anyone modeling 'real world' data to get an understanding of the sensitivity across different thermochronometric mineral systems with respect to a proposal history.
We agree that any individual can model sensitivity of real world data. The challenge is that with the growing user base of thermochronology, many people have not been trained to conduct sensitivity analysis. As we state in the text, benchmark paths can be used for a variety of applications because they are designed to highlight the sensitivity of each thermochronometric system to differences in kinetics models, tT history features, mineral compositions/geometries, and other variables critical to the interpretation of thermochronologic data. Introducing benchmark paths is a practical foundation for sensitivity analysis that is a tool for understanding the sensitivity of real world data.
Citation: https://doi.org/10.5194/gchron-2024-20-AC1 -
AC3: 'Statement on Goals of the Manuscript', Andrea Stevens Goddard, 17 Dec 2024
This manuscript was designed specifically for the Technical Note format and goals as described by GChron. It is not a research paper, but proposes how a new tool, benchmark paths, can be used to better understand and quantify the impacts of the decisions inherent in modeling and interpreting thermochronologic data. As part of our contribution we describe a process of tuning benchmark paths which permits practical, quantitative comparisons of the tT history differences that arise from modifying variables including, but not limited to, kinetics models, tT history features, mineral compositions/geometries. We demonstrate this tuning approach using different chronometer kinetics models.
This contribution provides formal guidance on modeling skills that are critical to robust data interpretations, but typically not formalized in research papers either in the main text or supplementary text. Peer-reviewed publication provides wide and equitable access to learning thermochronologic data analysis and interpretation techniques. The content is aligned with both the solicitation for Technical Notes by GChron and the Special Issue to which it was submitted, “Technical notes on modelling thermochronologic data”. Moreover, it is written to address a community need for formalized documentation and training in modeling techniques that has been part of a recurring community conversation at International Meetings (e.g. International Conference on Thermochronology, 2021 & 2023). A key audience for this publication is individuals within and outside the thermochronology community interested in growing their technical skills in modeling and data interpretation.
Citation: https://doi.org/10.5194/gchron-2024-20-AC3
-
AC1: 'Reply on RC1', Andrea Stevens Goddard, 17 Dec 2024
-
RC2: 'Comment on gchron-2024-20', Anonymous Referee #2, 18 Oct 2024
Why do we need new benchmarks? I don’t think this short paper has made a strong case. Tellingly, in the discussion (lines 227-232) authors note that ..”a single suite of tuned paths cannot capture all complexities of these systems”……. Very true so why are these new benchmarks fit for purpose?
We know models are imperfect and will continue to be improved upon. Yes, it is important to test and compare different models to determine their efficacy but this needs to be done in depth exploring the full range of predictions for a wide range of geological timescales, compositions and thermal history complexity. This study is a mere snapshot.
The greatest challenge with models is the extrapolation from laboratory to geological timescales. We can only know if a model works through comparison with geological data and this is why researchers were motivated to compare different (U-Th)/He and FT models with data from borehole samples with a well constrained independent geological/thermal history which is what Naeser, (1979, 1980, 1981); Green et al., (1989) and later, House et al., (2002) attempted to do.
The discussion highlights some of the many limitations/uncertainties of (U-Th)/He and FT models that undermine the case for benchmark paths because how do you know your benchmark is a sensible reference point given the dispersion of predictions present in other models. Perhaps your benchmarks are outliers.
A final point regarding experimental design. Why choose a single composition for both AFT and AHe? – Are they representative of the most common types of apatite? The choice of using DPAR is questionable in that a DPAR of 2.05 µm can be found in apatites with between 0.02 wt% Chlorine and up to 0.8 wt% chlorine. Model outcomes would very different for some thermal histories has chlorine been used instead of DPAR.
Citation: https://doi.org/10.5194/gchron-2024-20-RC2 -
AC2: 'Reply on RC2', Andrea Stevens Goddard, 17 Dec 2024
In our reply, we copied referee comments for reference and have replied to each of those comments in italics below.
Anonymous Referee #2
Why do we need new benchmarks? I don’t think this short paper has made a strong case. Tellingly, in the discussion (lines 227-232) authors note that ..”a single suite of tuned paths cannot capture all complexities of these systems”……. Very true so why are these new benchmarks fit for purpose?
We know models are imperfect and will continue to be improved upon. Yes, it is important to test and compare different models to determine their efficacy but this needs to be done in depth exploring the full range of predictions for a wide range of geological timescales, compositions and thermal history complexity. This study is a mere snapshot.
We want to point out that the intention of this Technical Note is not to determine the efficacy of kinematics models. Our intention is to develop a new tool, benchmark paths, to explore how different inputs in thermal history modeling should be considered in data interpretations. As an example, we use benchmark paths to explore the sensitivity of kinetics models to different types of thermal histories.
We agree that this is a snapshot, but importantly it introduces a new tool for comparing different models that we anticipate will be incorporated into texts introducing new kinetics models and also lays the foundations for the personal training and practice of thermochronologists to quantify differences among kinetics models themselves. We intentionally describe methods of tuning benchmark paths to explore these sensitivities so that this topic can be expanded formally or informally. A single, exhaustive work on this topic would be impractical and inefficient.
The greatest challenge with models is the extrapolation from laboratory to geological timescales. We can only know if a model works through comparison with geological data and this is why researchers were motivated to compare different (U-Th)/He and FT models with data from borehole samples with a well constrained independent geological/thermal history which is what Naeser, (1979, 1980, 1981); Green et al., (1989) and later, House et al., (2002) attempted to do.
We agree. The purpose of benchmark paths is not to establish the efficacy of a particular model, but to quantify (in an intuitive way) differences among models. Geological tests of models are incredibly valuable, but beyond the purpose and scope of this Technical Note.
The discussion highlights some of the many limitations/uncertainties of (U-Th)/He and FT models that undermine the case for benchmark paths because how do you know your benchmark is a sensible reference point given the dispersion of predictions present in other models. Perhaps your benchmarks are outliers.
Whether a benchmark path is a sensible reference point will depend upon the geologic problem. The goal of this paper was to provide a strategic subset of benchmark paths that help modelers get a quantitative grasp of how thermochronologic data can be sensitive to differences in kinetics models, tT history features, mineral compositions/geometries, and other variables. In some cases benchmark paths may provide a sufficient understanding for a modeler to move forward. In other scenarios, a modeler may need to expand using different paths to explore sensitivities unique to their particular study. Importantly, this manuscript (1) formalizes the importance of exploring these questions, (2) provides a new tool, benchmark paths, which can be used by thermochronologists to explore uncertainties particularly relevant to their own data set, and (3) provides examples of how to use and tune benchmark paths.
A final point regarding experimental design. Why choose a single composition for both AFT and AHe? – Are they representative of the most common types of apatite? The choice of using DPAR is questionable in that a DPAR of 2.05 µm can be found in apatites with between 0.02 wt% Chlorine and up to 0.8 wt% chlorine. Model outcomes would very different for some thermal histories has chlorine been used instead of DPAR.
We agree! This is an additional sensitivity that can be explored, but was beyond the scope of this manuscript. As the coordinators of this Special Issue we anticipate that there are authors preparing demonstrations highlighting this exact point which we think is a very practical demonstration, particularly relevant for this Special Issue.
Citation: https://doi.org/10.5194/gchron-2024-20-AC2
-
AC2: 'Reply on RC2', Andrea Stevens Goddard, 17 Dec 2024
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 197 | 77 | 121 | 395 | 12 | 12 |
- HTML: 197
- PDF: 77
- XML: 121
- Total: 395
- BibTeX: 12
- EndNote: 12
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1